What is the description of the source database restriction type in RDS MySQL synchronization to polardb MySQL Cluster?

The table to be synchronized must have a primary key or unique constraint , And the field is unique , Otherwise, duplicate data may appear in the target database . For example, the synchronization object is table level , And need to edit （ Such as table column name mapping ）, Then a single synchronization task only supports synchronization 1000 A watch . When the quantity limit is exceeded , The request error will be displayed after the task is submitted , At this time, it is recommended that you split the table to be synchronized , Configure multiple tasks in batches , Or configure the synchronization task of the whole library . Binlog journal ： To be opened Binlog, For the opening method, see setting instance parameters , also binlog_format by row、binlog_row_image by full. Otherwise, an error will be prompted in the pre inspection stage , And the data synchronization task cannot be started successfully . Be careful If the source instance is self built MySQL Dual master cluster （ The two are master and slave to each other ）, To protect DTS Can get all Binlog journal , Then you need to turn on the parameter log_slave_updates. If it is an incremental synchronization task ,DTS The local address of the source database is required Binlog Log saving 24 hours , For example, full synchronization and incremental synchronization tasks ,DTS The local address of the source database is required Binlog Keep the log at least 7 Days or more （ After the full synchronization is completed, you can Binlog Save time is set to 24 hours ）, otherwise DTS May be due to inability to obtain Binlog And cause the task to fail , In extreme cases, it may even lead to data inconsistency or loss . Because of the Binlog Log save time is less than DTS The required time leads to problems , be not in DTS Of SLA Within the scope of protection .